SfTI invests in 15 new Seed projects to the total of $2.97m

22 June 2020 | Read time: 21 minutes

Hero image for news page Seed project fund announcement Image770x400

The 2020 Seed Projects funding call closed on 20 March, 2020 with 108 proposals received. The assessment panel carefully assessed each application which resulted in 15 successful projects. Read on for full list of the 2020 Seed projects and learn more about the exciting breakthroughs and discoveries ahead.

About the Seed Project fund

Funded for up to two years with a maximum of $200,000 per project. The Seed project fund is intended to bring in new ideas and new researchers to the SfTI community. Our Seed projects are smaller, high risk, technically complex, projects based on ideas developed by researchers themselves. From 2016-19 SfTI has invested in 54 Seed projects, with funding rounds typically held every year. Learn more about the projects we have funded.

Proposal theme alignment

The SfTI Challenge is based upon four areas of capability, or Themes. These are broad, multidisciplinary research topics that potentially span many different applications and industry sectors. The 2020 funding round gave priority to proposals:

  • aligned with the Data Science and Data Technologies (DSDT) or Sensors, Robotics & Automation (SRA) themes
  • with a team (including the Principal Investigator) of emerging researchers and;
  • Vision Mātauranga (VM) aligned research.

Of the 108 proposals received, 71 claimed Data Science and Digital Technologies (DSDT) theme alignment (primary or secondary), 67 claimed Sensors, Robotics and Automation (SRA) theme alignment, and 34 claimed Materials, Manufacturing Technologies and Design (MMTD) theme alignment.

Of the recommended proposals, 12 of 15 are aligned with the DSDT theme (primary or plus secondary), 8 of 15 aligned with the SRA theme, while MMTD alignment was 3 of 15. This is consistent with SfTI’s priority for DSDT and SRA proposals in this round, to balance our portfolio of research projects across our technical themes.

Thirty-four (34) proposals claimed alignment with the Vision Mātauranga (VM) theme and seven (7) of these were successful including five (5) with very strong VM alignment. 

Our Vision Mātauranga process

Our Vision Mātauranga Seed projects undergo a rigorous process to ensure that Māori aspirations are met. All projects strongly encourage Māori leadership where capacity and capability is available and where it isn’t, Māori team members, experts and community involvement in the design, development, research and output phases are essential. Vision Mātauranga projects are carefully reviewed by leading Māori experts in the field and when implemented are overseen by an experienced Māori academic mentor.

Capacity and capability of Māori in the technology and engineering space are a key area that SfTI is dedicated to supporting. We understand the under-representation of Māori in these areas. To improve representation, SfTI encourages collaboration and co-creation between researchers in the field and upcoming Māori researchers, Māori communities and organisations that will lead to future projects with Māori leadership.  We strongly encourage projects that are Māori-led and would be more than happy to talk to anyone who is keen to develop a proposal.

Vision Mātauranga project descriptions

Here we have expanded on our Vision Mātauranga Projects and describe how our Māori researchers, community and organisations are involved. We have also clarified which projects were funded via our Vision Mātauranga pool and which ones we are encouraging to further develop their Vision Mātauranga. Māori researchers are not publicly named here yet, as we are seeking permission to do so. We appreciate your patience.

Successful 2020 Seed projects

Practical solutions for outdoor localisation using acoustics in conservation

Dr Yusuke Hioka and Professor Stephen Marsland (VUW)

University of Auckland, $199,996

Themes: Data science and digital technologies / Sensors, robotics and automation

Monitoring New Zealand’s native species is particularly important for conservation. The primary aim of this project is to apply our previous work on sound localisation in natural environments into a useful tool for citizen-led conservation projects. The research will first establish a reliable algorithm for accurate sound source localisation in outdoor environments by extending the algorithms we developed in the previous SfTI project. This algorithm will then be applied to three different practical applications: estimation of bird abundance from their calls, localising incubation burrows of North Island brown kiwi, and localising Matuku/bitterns from the male booms. The research will ultimately lead to improved monitoring procedures of endangered NZ avian species.

Although not funded by our Vision Mātauranga Pool, this project has good alignment to Vision Mātauranga. This project builds and relies on previous relationships and collaboration with local iwi, the Māori arm of DOC and other Māori organisations.  The potential here are the benefits to the health and wellbeing of taonga species and forests. Whereby this project not only estimates bird abundance and localisation but also explores the presence of predators by measuring non-bird related sounds in the forest.

Automating automated testing

Dr Michael Homer

Victoria University of Wellington, $199,287

Theme: Data science and digital technologies

The growth of software in society has led to challenges and profound changes in the field of software engineering, such as the rise of agile methodologies that have revolutionised software engineering in the last 20 years. We propose to devise a novel method to automatically synthesise iteration tests to accurately describe faults from error stack traces and logs. Tests reproducing faults not only safeguard product releases, but also help developers to debug systems. If successful, the outcomes of this project will significantly increase the velocity and reliability of agile processes.

Novel approach to detect and predict tests with non-deterministic behaviour

Dr Amjed Tahir

Massey University, $200,000

Theme: Data science and digital technologies

Society relies on the proper functioning of software systems. These systems can be disrupted by software defects, resulting in catastrophic failures, for example the recent Boeing 737 MAX and Airbus A350 bugs. One of the methods to ensure software quality and reduce defects is software testing. Developers rely on tests to expose defects early when they are relatively less expensive to fix. Ideally tests result in a pass or fail. Otherwise they are known as flaky tests which can undermine the usefulness of testing. Detecting flaky tests will help to intercept defects earlier and will improve the overall quality of software programs. In this project, we aim to understand the causes of flakiness, provide empirical evidence of new patterns of flaky tests and their classifications. We will then develop a novel technique that is able to identify and predict them with a high-level of accuracy and scalable to large programs.

Sensitive soft robotic skin for handling delicate objects

Dr Samuel Rosset

University of Auckland, $199,184

Themes: Sensors, robotics and automation

Safe interaction between conventional hard robots and delicate objects remains a challenge. This limits the use of robotics in fields such as healthcare or mechanised fruit harvesting. We plan to develop a soft compressible sensing skin that will provide robotic manipulators with the ability to detect and control the amount of force they apply while also providing a compliant and deformable interface with the object they are manipulating. Current literature reports that sensing skins usually rely on a variety of stretchable electronic components and complex structuring of compliant materials, making them costly and difficult to implement. We will develop an ultra-soft capacitive touch sensitive sensor with a monolithic structure and a single pair of electrodes that can inform on both the location and magnitude of an applied force.

Although not funded by our Vision Mātauranga Pool, this project has been recognised as having Vision Mātauranga growth potential that would be beneficial to Māori business and thus contributing to the Māori economy.

Physical activity promotion chatbot for youth wellbeing

Dr Sarah Hopkins

University of Auckland, $200,000

Themes: Data science and digital technologies / Sensors, robotics and automation / Vision Mātauranga (Funded)

Being physically active is linked to increased wellbeing. This study will co-design a chatbot to help young people (11-19 years of age) to identify and persist with physical activity. The tool will be co-designed with Māori and Pasifika youth to develop a culturally relevant and effective digital mental and physical health tool using a holistic approach consistent with the Whare Tapa Whā model. The co-design process will be guided by the Community-Up Research Practices to ensure that rangatahi Māori voices are heard, that whānau can be partners and protected in the research. These processes will be guided by Māori research experts, including Tania Cargo (Ngāti Maru, Ngāti Manu, Ngāpuhi), a clinical psychologist and high performance sports coach who is a key member of our research team.

The work will be integrated with ‘Health Advances through Behavioural Intervention Technologies (HABITs), a programme developed from the ‘A Better Start - E Tipu E Rea’ National Science Challenge. The chatbot will include machine learning of dialog policy to identify user preferences and opportunities and provide tailored and appropriate physical activity recommendations; and sequence learning to provide naturalness in conversational responses. This will allow better personalisation and longer engagement than purely template-based methods. 

Wearable physiological sensors for head mounted displays

Dr Amit Barde

University of Auckland, $200,000

Themes: Sensors, robotics and automation / Data science and digital technologies

This project will develop Electroencephalography (EEG), Electromyography (EMG) and Electrooculography (EOG) sensors based on conductive fabric that will be more comfortable to wear over long periods of time, in comparison to existing technology. We will also push the boundaries to integrate multiple sensors in head mounted displays (HMDs) that will create a more accurate understanding of user behaviour and intention. Our work targets physiotherapy and facial stroke rehabilitation, however, we anticipate this technology extend to a range of industry and research applications.

Surrogate modelling with wearable sensors to estimate joint torques, muscle forces and beyond

Dr Ted Yeung

University of Auckland, $199,333

Themes: Sensors, robotics and automation / Data science and digital technologies

The loss of mobility of the lower limb in patients with movement disorder can affect quality of life. Understanding human gait and its effect on the surrounding soft tissue is essential to assess and monitor abnormal human motion. Traditional optical motion capture methods collect gait data in a controlled environment that only provides small and restrictive snapshot of daily living. With the advent of wearable sensors and smartphones, we can collect human gait data outside of the clinic into the community. Our aim is to develop a novel machine learning workflow to create custom surrogate models that use wearable sensors to predict human gait properties in the real world.

Ngā taonga o nehe rā me te heke mai: treasures of the past and future

Professor Steve Reeves

University of Waikato, $199,992

Themes: Data science and digital technologies / Vision Mātauranga 

The preservation, authentication and transfer of mātauranga and traditional knowledge for Māori and indigenous populations is an area in need of urgent innovation. Everything from languages, family histories and local narratives to traditional medicines, all are at some risk of being lost due to the modern day challenges of maintaining traditional practises of knowledge management. However, the current technical mechanisms for storing, managing and sharing data and information does not fit with how traditional communities protect and cultivate their knowledge. This project seeks to use and develop technologies that will mirror how Māori knowledge is best maintained. It has to be secure, to maintain data sovereignty; trusted to maintain authenticity; decentralised so that the community can express their own governance traditions in its management; and sharable, so that the community can use it to collaborate when they want to do so. Working with Māori experts from Āhau NZ Limited, and with technologies including, Blockchain and Tangles this project aims to create an open-source prototype tool that meets these aspects to allow communities to use technology to maintain and grow their Mātauranga.

This project was not funded by the Vision Mātauranga pool. However, we see that this project has strong Vision Mātauranga alignment working with Āhau NZ Ltd and SfTI will provide Māori expert mentorship throughout the project to further develop their Vision Mātauranga capability and widen their collaboration with Māori communities.

Stretchable and self-healable energy storage for epidermal and implantable bioelectronics

Dr Bicheng Zhu

University of Auckland, $200,000

Themes: Materials, manufacturing technology and design / Sensors, robotics and automation

With the advent of epidermal and tissue-implantable bioelectronic devices, there is an urgent need to develop wearable energy storage solutions, especially for continuous health monitoring and medical devices. Supercapacitor-based energy storage systems are a promising technology with advantages such as high-power density, long cycle life and simple device fabrication. We plan to extend this technology to develop stretchable, self-healable and biocompatible conjugated polymer-based supercapacitors for use in medical devices and continuous health monitoring electronics.

Smart surgical planning and guidance system for anatomical ACL reconstruction.

Dr Marco Schneider

University of Auckland, $200,000

Themes: Data science and digital technologies / Sensors, robotics and automation/ Materials, manufacturing technology and design

Approximately 40% of anterior cruciate ligament (ACL) grafts are misplaced as a result of reconstructions performed by traditional arthroscopy. We plan to develop a smart preoperative planning software system for defining anatomically accurate graft tunnel locations in ACL reconstruction surgery using machine learning and musculoskeletal models; and to develop a multimodal 3D motion tracked intraoperative system to register and guide the surgeon’s tools to the correct location in the ruptured knee. This system will enable ACL reconstruction surgeries to be performed more reproducibly and accurately. 

Domain adaptation to support Polynesian language technologies

Dr Jonathan Dunn

University of Canterbury, $200,000

Themes: Data science and digital technologies / Vision Mātauranga (Funded)

Recent work has shown that just 20 languages account for 90% of all online language use. Without the support of language technologies, minority languages will never thrive in digital environments. Technologies such as information retrieval require increasingly large amounts of training data, but there is relatively little data available for Polynesian languages like Māori and Tongan. This project enables the development of language technologies for low-resource Polynesian languages by creating novel domain adaptation methods. Working with Professor Jeanette King from the University of Canterbury, Māori, non-Māori and Pacific linguists will contribute to the development of this project that will enable Māori and other Pacific language speakers to use their own language in more situations in a digital space.  Domain adaptation is the only approach capable of bridging this gap to bring an inclusive multi-lingual digital environment to Aotearoa New Zealand and contribute to enhancing language experience for Māori and Pacific peoples with such technologies as grammar check, question answering systems, digital assistants and more.

Unlocking Mātauranga Māori in historic survey maps

Dr Stuart Mead

Massey University, $197,840

Themes: Data science and digital technologies / Vision Mātauranga (Funded)

Surveyors preparing early (1800’s) land survey maps in New Zealand relied on the assistance and mātauranga Māori of local expert guides, and often embedded significant information on cultural sites, place names and the environment within hand-drawn maps and notes. The mātauranga Māori embedded within historic maps is largely under-utilised as the data are ‘locked’ in machine unreadable formats and unavailable for analysis. We will develop computer vision and geospatial techniques to extract and geo-locate information in survey plans, unlocking and transforming this Māori knowledge into feature-rich, digital geospatial information. The project conception and co-creation is done with Māori geospatial practitioners from within Te Kahui Manu Hokai (Maori GIS Association) to meet the need to secure Māori access to and analysis of this data. Thus this has created a deeper  understanding of management, use and needs of Māori geographical information, whilst ensuring that principles of data sovereignty are fulfilled as outlined in Te Mana Raraunga Brief. Mentorship from a senior Māori academic is a critical component of this project alongside the development of capability of all researchers involved as well as capacity development of Māori leadership through the mentoring of a Māori PhD student.  The future potential use of this data spans a wide range of possibilities that can provide information on historical context on landscapes and land-use change, which has the potential to enhance Māori decision making surrounding economic and environmental development.

Proofing code reuse: Detecting errors in publicly available software

Dr Sherlock Licorish

University of Otago, $200,000

Theme: Data science and digital technologies

Software developers are increasingly dependent on question and answer portals and blogs for coding solutions. While such interfaces provide useful information, there are concerns that code snippets hosted here are often of poor quality, which may result in software failures. The objective of this project is to develop an appropriate code quality model and build an exhaustive repository of popular code snippets from online portals and blogs, to help predict the quality of these snippets for software developers.

Gait training using smart insoles and immersive reality technologies (XR)

Dr Rafael Kuffner Dos Anjos

Victoria University of Wellington, $200,000

Themes: Data science and digital technologies / Sensors, robotics and automation

Gait analysis and rehabilitation training are areas of healthcare which dramatically increase the quality of life of post-trauma patients, athletes, people with chronic conditions and many others. Traditionally these tasks have been done through movement observation by a medical doctor. More recently, the development of tracking technology and wearables such as smart insoles and easily accessible AR/VR devices have opened new opportunities for gait training and rehabilitation. This project aims to leverage state-of-the-art immersive reality (XR) and sensing technologies (Smart Insoles) to develop novel portable real-time tools and new knowledge to  transform how gait training is currently performed for rehabilitation and physical recovery.

Weaving Maori culture into natural fibre reinforced composites

Dr Tim Huber

University of Canterbury, $173,879

Themes: Materials, manufacturing technology and design / Vision Mātauranga (Funded)

Cultivars of Harakeke (Phormium tenax and Phormium colensoi) or New Zealand flax, are woven by Māori weavers into patterns often unique to the individual weaver, region and iwi in the form of tukutuku, kete, and fishing nets. These patterns have emerged as a combination of art, spirituality and functionality. Modern composites using natural fibres are attractive alternatives to non-sustainable composites, but a lack of suitable woven textiles has been identified as a major barrier for the use of natural fibres in high performance composites. This project involves a multidisciplinary approach and co-design with Māori weavers and experts. Our team including a key Māori expert weaver will analyse Māori weaving patterns and the traditionally used Harakeke cultivars to identify weave patterns and Harakeke cultivars likely to provide high strength/stiffness in a composite.  This work is done collaboratively with iwi and contributes to Māori academic success through capacity development and the design and implementation of Māori educational resource design for pāngarau (mathematics) using ethnomathematics.